pillar

Gemini vs Perplexity Traffic: Raw Clicks vs. Real Value

AI referral traffic is growing 357% year-over-year, but not all AI clicks are created equal.

Gemini Flipped the Traffic Scoreboard in Four Months

In August 2025, Perplexity was sending roughly three times more referral traffic to websites than Google Gemini. By January 2026, Gemini had overtaken it, sending 29% more visitors globally and 41% more in the United States. That reversal happened in about four months.

Most of the coverage focuses on the volume story. Gemini is bigger now, Perplexity is smaller, ChatGPT still dominates everything. Fine. But the deeper I dig into this data, the more I think the raw traffic numbers are answering the wrong question. The comparison that actually matters is which of these AI search engines sends clicks worth having.

The answer is not what the headline numbers suggest.

The Gemini growth numbers are genuinely striking. SE Ranking's analysis of 101,574 websites shows Gemini's referral traffic grew 388% between September and November 2025, then essentially doubled again through January 2026. Similarweb data puts the year-over-year growth at 643% as of February 2026.

The timing isn't coincidental. Google rolled out Gemini 3 between November 18 and December 17, 2025. The model upgrade seems to have changed how often Gemini includes outbound links in its answers. December referral traffic jumped 51.5%, January added another 42%.

Meanwhile, ChatGPT's referral traffic dropped 22% from October 2025 to January 2026. It still dominates the category. ChatGPT drives roughly 87% of all AI referral traffic and holds about 64.5% of generative AI web traffic share. But the gap is narrowing. ChatGPT's lead over Gemini shrank from 22x in October to 8x by January. If Gemini maintains its trajectory (which, to be fair, is a big if), SE Ranking models suggest it could match ChatGPT's referral output sometime in late 2027.

Perplexity grew too. Referral traffic from Perplexity grew 490% year-over-year across over 7,000 US websites, according to Semrush data reported by Adweek. It just didn't grow fast enough to keep pace with Gemini's surge.

So if you're looking at the volume scoreboard: ChatGPT first, Gemini second, Perplexity third. Simple enough.

The Part About Gemini vs Perplexity Traffic Nobody Is Comparing

This is where the comparison gets more interesting, and honestly where I think most of the analysis falls short.

Seer Interactive published a conversion study across AI search platforms. The numbers:

ChatGPT: 15.9% conversion rate
Perplexity: 10.5%
Claude: 5%
Gemini: 3%
Google Organic: 1.76%

Read that list again. Perplexity's conversion rate is 3.5 times higher than Gemini's. ChatGPT's is more than 5x higher. Every AI platform converts better than traditional organic search, but the spread between them is enormous.

Ahrefs found something similar from their own data: AI search visitors accounted for just 0.5% of their total traffic but drove 12.1% of signups. That's a 23x conversion multiplier. AI visitors also viewed 50% more pages per session and bounced less often.

This creates a genuinely awkward math problem. Gemini sends more raw visits than Perplexity. But if Perplexity visitors convert at 3.5x the rate, Perplexity might be delivering more actual value despite the lower volume. For a site that cares about signups, purchases, or leads (which is, you know, most sites), the Perplexity visitors could be worth more in aggregate even with fewer of them arriving.

Why the conversion gap? I think citation behavior explains a lot of it. Perplexity cites its sources in 97% of responses. ChatGPT cites sources about 16% of the time. Google AI Overviews land around 34%.

When Perplexity links to your site, it's usually in a context where the user specifically needs the kind of detail your page provides. The click is intentional. The user has already been primed by the AI summary and is going deeper on purpose. When Gemini sends traffic, the intent signal is fuzzier. It's a bigger, more general-purpose product with 650 million monthly active users, and many of those outbound clicks are more casual.

There's also the crawl-to-refer ratio, which is probably the most underreported metric in this entire conversation. SEOmator analyzed how many pages each AI bot crawls for every one referral visit it sends back:

PerplexityBot: 111 pages crawled per referral
GPTBot (OpenAI): 1,276 pages per referral
ClaudeBot (Anthropic): 23,951 pages per referral
Microsoft Copilot: 33 pages per referral

For comparison, Google's ratio is about 4.7:1 and DuckDuckGo is 1.5:1.

Perplexity is, page for page, far more efficient at turning crawls into actual referrals than any other AI bot. If you're deciding which bots to allow through your robots.txt (a decision we touched on when Google launched its agent crawler), this ratio matters. You're trading server resources for traffic. Perplexity's trade is roughly 11x more efficient than OpenAI's and 215x more efficient than Anthropic's.

Get NMS in your inbox. Actionable marketing intelligence, Mon-Fri. No fluff, no puns, just the stuff that moves numbers. Subscribe free.

The Number That Reframes the Entire Gemini vs Perplexity Debate

Most AI traffic articles leave out an uncomfortable detail: all AI referral traffic combined accounts for about 1.08% of total website traffic. Google traditional search still sends roughly 345 times more traffic than every AI chatbot put together.

The bigger disruption isn't coming from AI chatbots sending visitors around. It's coming from Google's own AI Overviews eating the clicks that Google Search used to send.

AI Overviews now appear in about 25% of Google searches. Where they show up, Ahrefs found a 58% reduction in click-through rate for the top-ranking page. That's up from 34.5% when they measured last April. Seer Interactive's data is even more stark: organic CTR fell 61% for informational queries with AI Overviews, and paid CTR dropped 68%.

Google AI Mode has a 93% zero-click rate, according to Semrush. Small publishers have lost 60% of their Google referral traffic over the past year. Medium-sized sites lost 47%. Even large publishers lost 22%.

To put it bluntly: Google is taking away far more traffic through AI Overviews than all AI chatbots are collectively sending back. The Gemini vs Perplexity comparison, while useful for people already thinking about AI search optimization, is a subplot inside a much larger story about the erosion of referral traffic from traditional search.

That doesn't mean AI referral traffic is irrelevant. The conversion quality data makes it genuinely interesting on a per-visit basis. But if you're spending hours optimizing for Gemini citations when you haven't audited your AI Overview exposure in Search Console yet, you're probably working on the wrong problem first.

Tracking AI Referral Traffic in GA4 Is Messier Than It Should Be

One practical problem with all of this: you might not know how much AI traffic you're actually getting. GA4 misclassifies a significant chunk of AI referral traffic as "Direct" or "Unassigned" because many AI platforms strip referrer headers or don't pass UTM parameters.

The fix is creating a custom channel group in GA4 with a regex that catches known AI referral domains. Something like:

Orbit Media has a solid walkthrough on the full setup. The whole process takes maybe 15 minutes and immediately gives you visibility into traffic that was previously invisible.

Two newer wrinkles worth knowing about: Perplexity launched Comet (a browser product) and ChatGPT launched Atlas (a similar interface). Traffic from these shows up differently in GA4 than traffic from the standard chat interfaces, and the referral signatures are still evolving. MarTech documented some of the current tracking behavior, but expect this to keep changing through 2026.

One more thing. If your AI referral traffic looks surprisingly low, check your robots.txt before assuming nobody's citing you. All three major publishers in Adweek's Perplexity study (Forbes, NYT, Guardian) had blocked PerplexityBot via robots.txt. They were still getting some referral traffic because Perplexity can cite pages it hasn't crawled directly (through search API access), but the volume was a fraction of what it could have been.

Which AI Platform to Prioritize (It Honestly Depends on What You Publish)

There's no universal "optimize for this one" recommendation here. It depends on your vertical and what matters to your business.

If you're in IT or technology publishing, AI referral traffic already accounts for 2.8% of total visits, well above the 1.08% average. Finance sites see very different Perplexity crawl behavior than shopping sites (a 42:1 crawl-to-refer ratio vs 182:1), which suggests Perplexity's algorithm treats verticals differently in how aggressively it links out.

For citation likelihood across all platforms, a few structural factors seem to matter more than anything on-page. Pages updated within the last two months earn about 28% more AI citations on average than pages older than two years. Pages with author schema markup get cited roughly three times more often. And the sources AI models lean on most heavily are Reddit and Wikipedia, not necessarily the most authoritative page on a given topic. ChatGPT cites Wikipedia in 48% of references, and Perplexity cites Reddit for 46.7%.

That Reddit stat is worth sitting with for a second. If you're creating content that gets discussed on Reddit, Perplexity in particular is more likely to surface and cite your work. Reddit's role in AI search distribution keeps growing, not shrinking. And as our analysis of 300,000 domains showed, some of the conventional wisdom about optimizing for AI visibility (like llms.txt files) isn't backed by the data.

For most marketing teams, I think the practical priority order looks roughly like this:

Audit your AI Overview exposure in Search Console. This is where the real volume impact lives.
Set up AI referral tracking in GA4 so you have actual data instead of guesses.
Review your robots.txt to make sure you're not blocking the bots you want traffic from.
Focus content optimization on whichever platform your vertical already gets the most traction from.

The Case That This Is All Premature

I should be honest about the limitations of the data here. The conversion rate numbers, while compelling, come from a relatively small number of studies. Seer Interactive's figures are based on their client base, which skews mid-market B2B. Ahrefs' conversion data is from their own product (an SEO tool), which attracts a specific kind of user. Whether those conversion rates hold across ecommerce, media, SaaS, and other verticals is genuinely unclear.

There's also the question of whether Gemini's traffic surge is durable or mostly a one-time bump from the Gemini 3 rollout. Google has a history of launching AI features with strong initial usage that settles down once the novelty fades. SE Ranking's data only covers a few months of post-surge traffic. The trendline could flatten.

And 1.08% of total traffic is still, well, 1.08%. For a site getting 100,000 monthly visits, that's roughly 1,080 from all AI platforms combined. Even with 4x conversion rates, you're looking at maybe 43 extra conversions per month from the entire AI referral channel. Worth monitoring. Probably not worth reorganizing your content strategy around unless you're already doing everything else right.

The trajectory is the part that's hard to dismiss, though. AI referral traffic grew 357% year-over-year. Analyst projections suggest AI search visitors could surpass traditional search visitors by 2028. The GEO market is projected to grow from $848 million to $33.7 billion by 2034. Those are growth curves where getting the tracking infrastructure in place early, even before the volume fully justifies it, tends to be the move that looks obvious in hindsight.

FAQ: AI Search Engine Referral Traffic

Does Gemini or Perplexity send more referral traffic to websites?

As of January 2026, Gemini sends about 29% more referral traffic than Perplexity globally (41% more in the US), according to SE Ranking's analysis of over 100,000 websites. This is a reversal from August 2025, when Perplexity was sending roughly three times more traffic. The shift coincided with Google's Gemini 3 model rollout in late 2025.

Which AI search engine has the highest conversion rate?

ChatGPT leads with a 15.9% conversion rate, followed by Perplexity at 10.5%, Claude at 5%, and Gemini at 3%, per Seer Interactive's study. All four convert significantly better than Google organic search at 1.76%. The conversion quality gap means Perplexity can deliver more total value than Gemini per-visit despite lower traffic volume.

Should I block AI crawlers like PerplexityBot or GPTBot?

Consider the crawl-to-refer ratio. PerplexityBot crawls 111 pages per referral sent, which is relatively efficient. GPTBot's ratio is 1,276:1, and ClaudeBot is 23,951:1. If server load from crawling is a concern, those ratios should inform which bots you allow. Blocking all of them means you're unlikely to get cited in those platforms' responses, which also means losing out on their higher-converting traffic.

How do I track AI referral traffic in Google Analytics 4?

GA4 often misclassifies AI traffic as "Direct" or "Unassigned." Create a custom channel group with a regex matching known AI domains (chatgpt.com, perplexity.ai, gemini.google.com, claude.ai, copilot.microsoft.com, etc.). Orbit Media and SALT.agency both have step-by-step walkthroughs. The setup takes about 15 minutes.

Like this kind of analysis? NMS delivers actionable marketing intelligence every weekday morning. Subscribe free and stop guessing what matters.